Sorry to double post, but to make this separate.  I posted 2 videos (one still processing) that shows some improvements for this circuit that should come in handy.

http://www.youtube.com/watch?v=FYL5P4V87Ok

*Fix for the negative bias against the base (potential failure point)
*Fix for the oscillation start up, now starts at 500mV
*Integration for frequency adjust with fine adjustments (raise and lower)
*Integration (in second video) for Capacitors to further adjust frequency (lower of course)

I think the pot connection is pretty key to guarantee an oscillation start and reduce risk of frying the B-E junction.

Thanks

Ya, sorry for the listing incorrectly there.  I started with .47 and and changed to .470ma, but then must have accidentally erased the zero instead of the decimal.
So it was 7.755w

Also, I was able to get 11 of the 7.5W LED bulbs and try getting more power out of the system.  I think everything's covered in the video, but max power was 29.2w for the 82w total bulbs connected (though not quite as bright as connected to 120vac). I'll have to wait until they get more bulbs in to see if I can run more and might need a bigger source;  but the transformer should take a lot, if I can drive it.
http://www.youtube.com/watch?v=-ncciyB6SKc

Thanks

The coil I'm using is from an old salvage out of a 50Kva Laser supply from a local university.  This is a setup transformer (88mh to 480mh), though was not a high power section.  Also the output wires are 600v rated so it was not HV either.  Also, it was 60hz driven and is iron core, so not designed for HF.  The core is a "loop" laminate design.  Secondaries coils are wrapped on either side and the primary is wound AROUND (EDIT 2: it is possible these are wound on one end of the bobbins after looking more) the outer of both coils.  It IS left in stock wind configuration, I haven't changed anything with it.  Just kept it around because I thought it may be useful someday.

I have another couple, more common transformers;  I'll see if I can get those to work also. 

I believe the biggest loss in this design is from the core and low Q.   Since, the only magical material to solve this is restricted to military (heh, metglas);  the optimal design may be difficult to find.
** If we can find ANY amorphous core with a Square BH we're in business!  I have found 3 manufacturers but only contacted Metglas so far, with the usual run-around.  Otherwise there are some square BH ferrites, and then it would be to finding the highest Al and permeability with the lowest core loss per cm2.

I have an E-core coming that's a bit larger, higher Al and more permeability;  with the same core loss per set (overall area) as Lasersaber's.  We'll see if that makes some difference. 

Ultimately though, it seems our max output may be related to the Energy stored in the secondary (primarily in Inductance) and the maximum flux we can impose on the primary (restricted by transistor power, Impedance, volts, etc).  I believe with the Bias resistor we can drive the primary up to our source or transistor Temp max.  So, then it would seem the trick is to attain as high of Q secondary as possible to maximize output. 

just rambling lol,
Thanks

OH EDIT:  The primary is center tapped but not used here.  If you measure the inductance of either half, you get 9mH;  however across the ends is 88mH.  So the primary winds are in opposition (as pointed out in the earlier LED vids with this trans.), least that's my take.  I may have to remove that covering sometime to better see its' construction.

One from Lidmotor _{Super Joule Ringer 2.0  -- My Replication}http://www.youtube.com/watch?v=o0fCQghOQ-E&feature=uploademail

Impressive! The transformer you are using is single output secondary?

from Lidmotor
_{Lasersaber Super Joule Ringer 2.0 --With multiple bulbs}http://www.youtube.com/watch?v=iVHfruW22PA&feature=uploademail

I love the simplicity of the Joule Ringer 2.0.  The usefulness of a small 12 volt powered circuit that can light multiple lights is a powerful breakthrough.  The fact that it self adjusts to load opens up the real possibility of home lighting run off 12 volts.  Very cool.

I wondered if the one transistor would work with an air core coil.  I do a lot of work with wind power and have been building controllers using the 2N2222 because of its non linear amplification.  The power in the wind is non-linear so this transistor type works well to capture available power in very low winds.
I built a small coil with a transistor/LED that powers a 14 watt CFL.  Very simple circuit.  At 6 volts the power consumed in the LED and CFL is less than a watt.  At 9 volts 2.5 watts.  I haven't tried 12 volts yet but my guess is that with a non-lenear amplifier, power consumed will be closer to 7 watts.  The 2N222 can handle .6 amps so it may work.  The Tesla type exciter doesn't draw more power as you add lights so the amp draw is limited.

Challenge:  see if you can improve and make this more efficient without adding components!

Thank you for your work and sharing

http://www.youtube.com/watch?v=pCbTVGYnWdc&list=UUGIblCTNadascluwDyRFobQ&index=1&feature=plcp

@Lynxsteam:

Nice lamp! That is how I imagined the use of a big air core coil in a lamp, it looks so retro.

I am impressed by the fact that the 2N222 does not become hot at 6V and 150 mA (I guess it also supports 9V at 250 mA for some time).

I just tried a LED (from the base of the transistor to ground) instead of the bias resistor (from the base of the transistor to the positive rail) and it did not work (besides, I am using the MJE13007 transistor). With my air core coils I seem to need the bias resistor. With the bias resistor the LED does not light up and does not make any difference.

Your trigger coil seems to do the trick. Nice idea to use three coils (primary from positive rail to collector of the transistor, trigger coil at the base and the big secondary just for the lamp). This circuit is so simple, but still there are endless variations.

Greetings, Conrad